CN116579269B - Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device - Google Patents
Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device Download PDFInfo
- Publication number
- CN116579269B CN116579269B CN202310849613.1A CN202310849613A CN116579269B CN 116579269 B CN116579269 B CN 116579269B CN 202310849613 A CN202310849613 A CN 202310849613A CN 116579269 B CN116579269 B CN 116579269B
- Authority
- CN
- China
- Prior art keywords
- grid
- wall
- processor
- unit
- boundary layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000011065 in-situ storage Methods 0.000 title claims abstract description 21
- 238000000354 decomposition reaction Methods 0.000 claims description 41
- 238000004364 calculation method Methods 0.000 claims description 37
- 238000004891 communication Methods 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 17
- 239000013598 vector Substances 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 7
- 239000012530 fluid Substances 0.000 abstract description 9
- 238000004088 simulation Methods 0.000 abstract description 2
- 239000010410 layer Substances 0.000 description 111
- 230000007704 transition Effects 0.000 description 29
- 238000010586 diagram Methods 0.000 description 15
- 230000008569 process Effects 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000010845 search algorithm Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 239000012790 adhesive layer Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000000470 constituent Substances 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- CIWBSHSKHKDKBQ-JLAZNSOCSA-N Ascorbic acid Chemical compound OC[C@H](O)[C@H]1OC(=O)C(O)=C1O CIWBSHSKHKDKBQ-JLAZNSOCSA-N 0.000 description 1
- 239000000853 adhesive Substances 0.000 description 1
- 230000001070 adhesive effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
- G06T17/205—Re-meshing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Algebra (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Fluid Mechanics (AREA)
- Computer Graphics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device, which relate to the technical field of engineering application of numerical simulation of multiple physical fields such as computational fluid mechanics and the like.
Description
Technical Field
The invention relates to the technical field of engineering application of numerical simulation of multiple physical fields such as computational fluid mechanics and the like, in particular to a method and a device for acquiring physical quantities of a polyhedral grid in-situ distributed parallel boundary layer.
Background
The flow area around the object in fluid mechanics can be divided into two areas for processing: one is a thin adhesive layer near the object plane, called boundary layer, the role of the adhesive in the boundary layer is very important, and the adhesive layer tends to have a strong shearing phenomenon; secondly, the regions outside the boundary layer are generally negligible in viscosity and are treated with non-viscous flow. Boundary layer theory proposed by Prandtl based on the above assumption has been called one of the basis stones of modern hydrodynamics. Boundary layer transition characterizes a continuous physical process of the boundary layer from a stable laminar state to a turbulent state, and is a leading edge problem in modern hydrodynamic research.
The boundary layer transition prediction method in the current computational fluid dynamics, such as a transition criterion method, a transition model method and the like, needs to acquire the value of the physical quantity in the boundary layer near the wall surface. The calculation efficiency and the calculation consumption of the main stream method are unacceptable, the calculation consumption can be reduced only by adopting sampling calculation of the boundary layer physical quantity under the condition of more iteration steps, and the boundary layer physical quantity is not updated in real time, so that the application of the transition mode and the transition criterion method in boundary layer transition prediction under a large-scale grid is greatly limited, the boundary layer transition prediction precision is reduced, the calculation cost is increased, and the boundary layer transition related engineering design period is prolonged.
Disclosure of Invention
The invention aims to provide a polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device, which solve the technical problems that boundary layer parameters are difficult to acquire with high efficiency and low memory in a computational fluid dynamics method based on the polyhedral grid, so that the application of a constructed transition mode and transition criterion method is limited, the computational cost is high and the like.
In a first aspect, the present invention provides a method for acquiring physical quantities of polyhedral grid in-situ distributed parallel boundary layers, including:
Broadcasting the wall grid unit of each processor to the rest processors, so that each processor establishes a k-d tree data structure of the wall grid unit based on the linear distance, and performs adjacent point search;
based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points, inserting the grid body unit number corresponding to each grid unit in each processor and the grid body unit of the overlapping area into a normal grid body unit linked list according to the normal distance;
each processor sequentially solves a non-viscous Euler (Euler) equation and a viscous Navier-Stokes equation according to a physical field solver to acquire flow field object understanding and update physical field information of a region decomposition overlapping region based on the physical field solution;
and each processor calculates boundary layer physical quantity characteristic quantity in the local processor according to the physical field information of the region decomposition overlapping region, and calculates global wall grid unit boundary layer characteristic physical quantity through a parallel calculation function global protocol.
With reference to the first aspect, the present invention provides a first possible implementation manner of the first aspect, wherein, before broadcasting the wall grid unit of each processor to the remaining processors, so that the each processor builds a k-d tree data structure of the wall grid unit based on the linear distance, and performs the step of searching for the neighboring points, the method further includes:
Initializing a parallel computing environment, carrying out distributed parallel region decomposition on an input grid, and creating a multi-layer overlapped parallel computing communication data structure.
With reference to the first aspect, the present invention provides a second possible implementation manner of the first aspect, wherein the step of broadcasting the wall grid unit of each processor to the remaining processors, so that each processor builds a k-d tree data structure of the wall grid unit based on the linear distance, and performs the adjacent point search includes:
each processor broadcasts the collected wall grid cell information to other processors; wherein the wall mesh unit information includes: the normal vector of the wall grid unit, the wall grid center coordinate, the grid point coordinates forming the wall grid and the global grid plane number;
according to the position of the processor where the wall surface is located after the region decomposition, local numbering is given to the grid unit of the global wall surface;
based on the linear distance, taking the face center coordinates of the global wall grid units as data, and establishing a k-d tree data structure for representing the wall grid units;
and determining the nearest point of the wall grid unit to the input point based on the nearest search of the input point of the k-d tree data structure.
With reference to the first aspect, the present invention provides a third possible implementation manner of the first aspect, wherein the step of inserting, according to a normal distance, a grid body unit number and an overlapping area grid body unit corresponding to each grid unit in each processor into a normal grid body unit linked list based on a nearest distance wall grid unit of the grid body unit determined by the adjacent point includes:
the following steps are repeatedly performed for each grid cell in each processor until each of the grid cells in each processor is traversed:
searching a nearest-distance wall grid unit corresponding to each grid body unit, and recording a grid surface number of the nearest-distance wall grid unit and a linear distance between the grid body unit and the nearest-distance wall grid unit;
according to the linear distance, solving the normal distance from the grid body unit to the wall grid unit, and inserting the grid body unit number into a normal grid body unit linked list of the wall grid unit according to the normal distance;
starting a communication interface, decomposing distance information of grid body units in an overlapping area by an exchange area, and arranging and inserting the grid body units in the overlapping area into a normal grid body unit linked list of a target wall grid unit corresponding to the grid body units in the overlapping area according to the normal distance.
With reference to the first aspect, the present invention provides a fourth possible implementation manner of the first aspect, where each processor sequentially solves a non-sticky euler equation and a sticky navistokes equation according to a physical field solver, and the step of obtaining a flow field object understanding and updating physical field information of a region decomposition overlapping region based on the physical solution of the flow field includes:
each processor calculates a physical field of a non-stick Euler equation of the local grid unit;
each processor calculates a viscous Navisk equation physical field for the local grid cell;
according to a parallel computing communication data structure which is pre-established with multiple layers of overlapping, determining non-sticky and sticky physical field variables of grid body units in an overlapping area decomposed by an interface exchange area;
and updating physical field information of the region decomposition overlapping region based on non-sticky and sticky physical field variables of the overlapping region grid body unit of the interface exchange region decomposition.
With reference to the first aspect, the present invention provides a fifth possible implementation manner of the first aspect, wherein each processor calculates a boundary layer physical quantity feature quantity in a local processor according to physical field information of the area decomposition overlapping area, and calculates a global wall grid cell boundary layer feature physical quantity by a parallel calculation function global protocol, including:
Repeating the following steps for each global wall grid cell in each of the processors until each global wall grid cell in each of the processors is traversed:
collecting non-sticky and sticky physical quantities in the flowing physical quantities according to a grid body cell linked list of the wall grid cells;
calculating the characteristic quantity of a local boundary layer according to the non-sticky physical quantity of a normal grid body unit of the wall surface grid unit;
determining characteristic points of each wall grid unit according to the local boundary layer characteristic quantity;
and utilizing a parallel computing function to globally reduce the first characteristic point of each wall grid unit, the first characteristic quantity and the normal distance of each wall grid unit, and reduce the second characteristic point of each wall grid unit, the second characteristic quantity and the normal distance of each wall grid unit.
With reference to the first aspect, the present invention provides a sixth possible implementation manner of the first aspect, wherein the method further includes:
calculating other boundary layer physical quantities of the wall grid units according to the global wall grid unit boundary layer characteristic physical quantity calculated by each processor and physical field information corresponding to a local processor;
And calculating a target physical field solving value according to the other boundary layer physical quantities.
In a second aspect, the present invention also provides a polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition device, including:
the search module broadcasts the wall grid unit of each processor to the rest processors so that each processor establishes a k-d tree data structure of the wall grid unit based on the linear distance and performs adjacent point search;
the interpolation module is used for inserting the grid body unit number corresponding to each grid unit in each processor and the grid body unit of the overlapping area into a normal grid body unit linked list according to the normal distance based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points;
the acquisition module is used for sequentially solving a non-sticky Euler equation and a sticky Navistos equation according to a physical field solver by each processor to acquire physical field information of a flow field object understanding and decomposing an overlapping area based on the flow field physical solution updating area;
and the calculation module is used for calculating boundary layer physical quantity characteristic quantities in the local processors according to the physical field information of the region decomposition overlapping region and calculating global wall grid unit boundary layer characteristic physical quantities through a parallel calculation function global protocol.
In a third aspect, the present invention provides an electronic device comprising a memory, a processor, the memory storing a computer program executable on the processor, the processor implementing the steps of the method of any of the preceding embodiments when the computer program is executed.
In a fourth aspect, the present invention provides a machine-readable storage medium storing machine-executable instructions which, when invoked and executed by a processor, cause the processor to implement the steps of the method of any of the preceding embodiments.
The embodiment of the invention has the following beneficial effects:
according to the method, boundary layer physical quantities corresponding to the wall grid units can be calculated in an in-situ distributed parallel mode, full-field data are not required to be copied in a moving mode, calculation efficiency is high, communication times and communication quantity are small, and memory occupation is small.
According to the boundary layer theory, the characteristic quantity 1 (first characteristic quantity) of the boundary layer is calculated and judged by using the non-viscous flow field and the viscous flow field, the characteristic physical quantity is simple and reliable to calculate, and the characteristic physical quantity judgment difficulty under the complex appearance can be effectively reduced.
And 3, the method uses the optimized k-d tree data structure to process the nearest point search problem, and the built k-d tree has high efficiency and is used as an independent module to be embedded into the existing program.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram showing steps of a method for acquiring physical quantities of polyhedral grid in-situ distributed parallel boundary layers according to the invention;
FIG. 2 is a schematic view of a two-dimensional airfoil polyhedral grid computing area corresponding to an input interface of the present invention;
FIG. 3 is a diagram illustrating decomposition and artificial boundaries of non-overlapping areas of map division according to the present invention;
FIG. 4 is a schematic diagram of a three-layer overlap region decomposition parallel computing region formed in the present invention;
FIG. 5 is a schematic diagram of each processor acquiring all wall grid cells according to the present invention;
FIG. 6 is a schematic diagram of the present invention after a k-d tree is constructed and inserted into a root node;
FIG. 7 is a schematic illustration of the wall normal distance as solved by the present invention;
FIG. 8 is a schematic diagram of a list of local wall normal units constructed in accordance with the present invention;
FIG. 9 is a schematic diagram of a partially exploded coordinate system used in calculating feature quantities in accordance with the present invention;
FIG. 10 is a schematic diagram of a local-global merge used in calculating feature quantities in accordance with the present invention;
FIG. 11 is a schematic representation of calculated boundary layer thicknesses of the present invention;
FIG. 12 is a schematic diagram of functional modules of a polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition device provided by the invention;
fig. 13 is a schematic diagram of a hardware architecture of an electronic device according to the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The transition of the boundary layer is a very common flow phenomenon, and has important influence on various flows in different fields of aviation aircrafts, turbomachinery, aerospace aircrafts and the like. For aviation aircraft flow, frictional resistance under turbulent flow of the aircraft is generally several times that under laminar flow, so improvement of sailing economy by laminar flow is expected, but in some cases, fast transition of the boundary layer is also expected, and further separation of the boundary layer flow is inhibited. For aviation turbine mechanical flow, boundary layer transition directly influences engine performance parameters such as total pressure recovery coefficient, propulsion efficiency and the like. The pneumatic heating problem of high-speed flight is very important for the aerospace vehicle, the boundary layer transition process is accompanied with the severe change of pneumatic heating, and the prediction of boundary layer transition is beneficial to the thermal protection design work of the aerospace vehicle.
Boundary layer transition prediction methods in computational fluid dynamics, such as transition criterion methods and transition model methods, need to acquire values of physical quantities in boundary layers near wall surfaces. For example, in transition prediction of compressible flow, the velocity profile of the boundary layer plays an important role, and the velocity component of the non-viscous flow outside the boundary layer, which occurs in the boundary layer and is perpendicular to the boundary layer, is called cross flow. The cross flow is comprehensively determined by the deflection angle of the non-viscous flow speed and the flow speed relative to the boundary layer, and is specifically expressed as follows: outside the boundary layer, the transverse flow speed is zero because the deflection angle is zero; since the flow velocity is zero at the wall surface, the cross flow velocity is also zero, and there is an inflection point in the cross flow velocity in the boundary layer. Because of the existence of the speed inflection point, the cross flow has instability, and the boundary layer transition is easy to be induced under the condition of certain external disturbance. Therefore, the physical quantity in the boundary layer has great predictive significance for transition of the boundary layer. In the existing computational fluid dynamics method, distributed computation is mostly performed based on a certain grid division, boundary layer grid units near a wall surface may not be in a local data structure, and it is very difficult to solve boundary layers on an in-situ data structure, so that a main acquisition method adopted by physical quantities in the boundary layer is to reconstruct a set of boundary layer normal grid data structure (or called normal integral path), copy and distribute the physical quantities in the whole computation area into a newly reconstructed data structure, and then perform boundary layer physical quantity computation in the newly established data structure. This process is not in-situ parallel-i.e. the original distributed parallel data structure of the physical field solver is not used, but rather a new global data structure dedicated to boundary layer physical quantity computation is reconstructed. Because the flow fields are not in-situ parallel, the whole process needs to be collected, copied and distributed for calculation, and the calculation needs to be concentrated and integrated into the original data structure again.
As described above, this calculation process has twice the necessary global serial operation, which is extremely large in traffic and memory occupation, since the typical computational grid of computational fluid dynamics is currently on the order of tens of millions to hundreds of millions. In actual calculation, the boundary layer physical quantity is often required to be frequently acquired, the calculation efficiency and calculation consumption of the current main stream method are unacceptable, the calculation consumption can be reduced only by adopting sampling calculation of the boundary layer physical quantity under the condition of more iteration steps at intervals, instead of updating the boundary layer physical quantity in real time, the application of the transition mode and the transition criterion method in boundary layer transition prediction under a large-scale grid is greatly limited, the boundary layer transition prediction precision is reduced, the calculation cost is increased, and the boundary layer transition related engineering design period is prolonged.
Based on the above, the polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method provided by the embodiment of the invention can acquire boundary layer parameters with high efficiency and low memory, so that the constructed transition mode and transition criterion method can be widely applied, and the calculation cost is reduced.
For the convenience of understanding the embodiment, the method for acquiring the physical quantity of the polyhedral grid in-situ distributed parallel boundary layer disclosed by the embodiment of the invention is described in detail, and can be applied to intelligent control equipment such as a server, an upper computer and the like.
FIG. 1 is a flowchart of a method for acquiring physical quantities of polyhedral grid in-situ distributed parallel boundary layers, which is provided by an embodiment of the invention.
Referring to fig. 1, the method mainly includes the steps of:
step S102, broadcasting the wall grid unit of each processor to the rest processors, so that each processor builds a k-d tree data structure of the wall grid unit based on the linear distance, and performs adjacent point search.
Step S104, based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points, the grid body unit number and the grid body unit of the overlapping area corresponding to each grid unit in each processor are inserted into a normal grid body unit linked list according to the normal distance.
And S106, each processor sequentially solves a non-sticky Euler equation and a sticky Navistos equation according to a physical field solver, acquires flow field object understanding and updates physical field information of the area decomposition overlapping area based on the physical solution of the flow field.
In step S108, each processor calculates boundary layer physical quantity characteristic quantity in the local processor according to the physical field information of the region decomposition overlapping region, and calculates global wall grid unit boundary layer characteristic physical quantity through a parallel calculation function global protocol.
Prior to step S102, the method further comprises step S101: initializing a parallel computing environment, carrying out distributed parallel region decomposition on an input grid, and creating a multi-layer overlapped parallel computing communication data structure.
The method comprises the steps of inputting a computational fluid dynamics computational region grid in a message passing interface (MessagePassingInterface, MPI) environment after initialization, preprocessing the grid by using a multi-layer region decomposition parallel computing method and system for the polyhedral grid, performing distributed parallel region decomposition on the input grid, creating a multi-layer overlapped parallel computing communication data structure, and considering the requirement of boundary layer physical quantity computation, wherein the overlapped layers of the region decomposition are typically three layers; FIG. 2 is a schematic view of a calculation region of a polyhedral grid of a two-dimensional airfoil, wherein the thick lines are airfoil walls; FIG. 3 is a schematic view of a non-overlapping computing area after the initial area decomposition, the computing area being decomposed into two parts for distribution to two processors, the wall parts of which are also decomposed into two parts; FIG. 4 is a schematic diagram of a grid of processors after creation of a multi-layered overlapping region-hierarchical parallel computing communication data structure, with shaded portions of the grid of overlapping region cells moved from other processors, the processors exchanging physical field data of the shaded portions via a common interface, as indicated by the arrows in FIG. 4.
In some embodiments, each processor may broadcast the wall grid cell of the present processor to all processors, each processor obtains a list of all wall grid cells of the computing area, builds a k-d tree (k-dimensional tree) data structure based on straight line distances for all grid cells in the list of all wall grid cells, and for closest point searching of the cells, as an example, the step S102 may include:
step 1.1), each processor broadcasts the collected wall grid cell information to other processors; wherein the wall mesh unit information includes: the normal vector of the wall grid unit, the wall grid center coordinate, the grid point coordinates forming the wall grid and the global grid plane number;
each processor collects normal vectors, wall grid surface center coordinates, grid point coordinates forming a wall grid and global grid surface numbers of local wall grid units, and broadcasts the collected wall grid unit information to global wall grid unit data storage vectors of all processors;
step 1.2), according to the position of a processor where the wall surface is located after the region decomposition, local numbering is given to the grid unit of the global wall surface;
Specifically, each processor unit assigns local numbers to the global wall grid units, the local number sequence is sequentially increased according to the wall grid sequence arrangement of each processor, the processor 0 is numbered firstly, and then the processor 1 is numbered on the basis of the processor 0; because the local numbers are arranged according to the processor, the actual processor position of the wall surface after the area decomposition can be rapidly determined according to the local numbers; fig. 5 is a schematic diagram of each processor after collecting wall information and numbering, and as can be seen from fig. 5, the same wall in each processor has the same number, for example, the wall numbers at the same positions of the processors on the left and right sides of fig. 5 are the same, the leftmost end wall surfaces of the processor 0 and the processor 1 are all numbered f10, and the rightmost end wall surface is all numbered f34; the wall grid units in the same processor have continuous numbers and are increased with the number of the processor, it can be understood that the wall grid units are continuously increased from f0 to f49, and only the wall grid unit numbers corresponding to a plurality of positions are shown in fig. 5;
step 1.3), based on the linear distance, taking the face center coordinates of the global wall grid unit as data, and establishing a k-d tree data structure for representing the wall grid unit; based on the k-d tree data structure, the nearest search of the input points determines the point in the wall grid cell that is closest to the input point.
Wherein, the face center coordinates of the global wall grid units are used as data to establish a k-d tree, and the k-d tree is a binary tree with each tree node being a k-dimensional point, and in the method of the invention, k is equal to 3. The k-d tree is established by inserting the face center coordinates of the wall grid units with the number wi point by point, and dividing the space into two half spaces by using the x-axis of the face center coordinates of the wall grid units wi as a division hyperplane. The subtree located to the left of the node at the left of the split hyperplane and the subtree located to the right of the node at the right of the split hyperplane. Fig. 6 is a schematic diagram of the spatial division after the insertion of the head node a, and the entire space is divided into two areas by the x-axis of the a point.
The sub-tree is then built up sequentially with the next face-centered coordinate y-axis inserted into the wall grid cell and the face-centered coordinate z-axis as the segmentation hyperplane. The k-d tree is used after its creation to find the nearest neighbor search of the input point to find the point in the k-d tree (i.e., in the wall grid cells) that is closest to the input point. The key parameter in the k-d tree building process is the maximum leaf node number maximum_leaf_number, distance function tolerance dist_eps, and grid cell bounding box size bounding_box. The maximum number of leaf nodes, maximum_leaf_number, determines the height and width of the k-d tree, and this parameter greatly affects the efficiency of the search algorithm since the search algorithm needs to traverse the height of the tree and then trace back, with an optimum value of 30.ltoreq.maximum_leaf_number.ltoreq.50 in the method of the present invention. The distance function tolerance dist_eps affects the accuracy of the nearest point search, and the calculation method is as follows: calculating the side lengths of all wall grid cells according to the grid point coordinates of the wall grid cells, wherein the minimum side length of all wall grid cells is resolution, and obtaining a distance function tolerance dist_eps by using dist_eps=resolution =beta, wherein 1e-8 beta is less than or equal to 1e-6, beta is an adjustable parameter, and represents scaling the length by a certain magnitude. The size of the bounding box of the grid cell is used for quickly positioning possible nearest points, and is a key for high efficiency of a search algorithm, and the method comprises a chord_min [ wi ] [ j ] and a chord_max [ wi ] [ j ], wherein wi is a wall grid cell number, and j is a coordinate axis number; the chord_min is the constituent grid point coordinate component infinitesimal of the wall grid cell, and the chord_max is the constituent grid point coordinate component infinitesimal of the wall grid cell. The grid cell bounding box size bounding box of the wall grid cell f10 is plotted in fig. 6.
In practical application, step S104 may be implemented by circulating all local grid cells in the local processor, searching the nearest-distance wall grid cell with the number wj corresponding to the grid cell with the number ci in the k-d tree by using a nearest-search algorithm, recording the number wj of the nearest-distance wall grid cell of the grid cell and the linear distance variable distance thereof, solving the normal distance_normal from the grid cell to the wall grid cell according to the linear distance, and simultaneously inserting the grid cell number into the normal grid cell linked list wall_normal_cell_list [ wj ] of the nearest-distance wall grid cell with the number wj according to the normal distance size; starting a communication interface, decomposing distance information of grid body units in an overlapping area by an exchange area, and inserting the grid body units in the overlapping area into a normal grid body unit linked list wall_norm_cell_list [ wj ] of a nearest-distance wall grid unit (target wall grid unit) with a corresponding number wj according to the normal distance size; as an alternative embodiment, this may be achieved by the steps comprising:
the following steps are repeatedly performed for each grid cell in each processor until each grid cell in each processor is traversed:
Step 2.1), searching a nearest-distance wall surface grid unit corresponding to each grid body unit, and recording a grid surface number of the nearest-distance wall surface grid unit and a linear distance between the grid body unit and the nearest-distance wall surface grid unit;
the method comprises the steps of circulating all local grid body units in a local processor, searching a nearest distance wall grid unit corresponding to the grid body unit with the number ci in a k-d tree by using a nearest searching algorithm, returning a searching result, namely the number wj of the nearest distance wall grid unit and a calculated linear distance variable distance, storing the linear distance variable distance in an array distance_cell [ ci ] which is also used for representing the linear distance, and storing the number wj in a nearest grid wall nearest_wall [ ci ]; wherein distance_cell [ ci ] represents that the distance of the ci-th grid cell is a value stored by a linear distance variable distance, and each grid cell has a value; the nearest-distance wall grid cell of the (ci) th grid cell is numbered wj, each grid cell has one nearest-distance wall grid cell, and the nearest grid wall nearest_wall [ ci ] formed by all nearest-distance wall grid cells.
Step 2.2), solving the normal distance from the grid body unit to the wall grid unit according to the linear distance, and inserting the grid body unit number into a normal grid body unit linked list of the wall grid unit according to the normal distance;
it should be noted that, since the distance calculated by the nearest point algorithm is the straight line distance from the grid cell center of the grid body with the number ci to the nearest distance from the grid cell center of the wall surface with the number wj, but not the normal distance, when the grid in the near wall area is obliquely distorted, the difference between the two is larger, but the result of taking the straight line distance in the far wall area is smoother, the following algorithm is adopted to calculate the normal distance: if the calculated linear distance_cell [ ci ]]Is greater than the feature distance L, normal distance_normal_cell [ ci ]]Directly take the linear distance_cell [ ci ]]Is a value of (2); if the calculated linear distance_cell [ ci ]]Is less than or equal to the characteristic distance L, normal distance_normal_cell [ ci ]]Equal to the linear distance_cell [ ci ]]Face normal to nearest wall grid cell numbered wj wall_norm [ wj ]]Inner product of (i), i.eThe method comprises the steps of carrying out a first treatment on the surface of the Fig. 7 is a schematic diagram of a straight line distance calculated for the grid body center of the grid body C and a calculated normal distance, and fig. 7 shows a schematic diagram of a head region grid body directly adopting the straight line distance as the normal distance; in fig. 7, M represents a dot at the end of a straight line, that is, a cell center of a grid to which the straight line points, and similarly, a cell center of a grid to which the straight line points is C.
Step 2.3), starting a communication interface, decomposing distance information of grid body units of the overlapping area by the exchange area, and arranging and inserting the grid body units of the overlapping area into a normal grid body unit linked list of a target wall grid unit corresponding to the grid body units of the overlapping area according to the normal distance.
Illustratively, inserting a grid body cell number ci into a normal grid body cell linked list wall_norm_cell_list [ wj ] of the nearest-distance wall grid cell with the number wj, wherein an algorithm is an insert sorting algorithm, a first linked list node larger than the value is found according to the normal distance_normal_cell [ ci ] as a key word of the insert sorting algorithm, and the cell number is inserted before the linked list node; fig. 10 shows that the normal grid body cell wall_norm_cell f22 corresponding to the wall f22, the processor 0 includes local grid body cells a, B, C, and the processor 1 includes local grid body cells D, E;
each processor starts the communication interface formed in the step S101, the exchange area decomposes the normal distance_normal_cell [ ci ] and the nearest grid wall face nearest_wall [ ci ] value of the overlapped area grid, inserts the overlapped area grid body units into the normal grid body unit linked list wall_norm_cell_list [ wj ] of the nearest grid wall face nearest to the nearest grid wall face grid unit wj formed by the nearest grid wall face grid unit wj with the corresponding serial number wj according to the normal distance size arrangement; the arrow indicator of fig. 8 shows the list of grid body cells normal to the area-resolved overlap region of the wall f22, and the processor 0 exchanges the grid body cells a, B, C for which the area-resolved overlap region is obtained; the processor 1 exchanges the mesh body cells D, E that have obtained the region-resolved overlap region.
As an alternative embodiment, each processor starts a physical field solver to sequentially solve a non-sticky euler equation and a sticky Navistos equation, acquire flow field object understanding, and update physical field information of a region decomposition overlapping region; this step S106 may include:
step 3.1), each processor calculates a physical field of a non-stick Euler equation of the local grid unit;
step 3.2), each processor calculates a viscous Navisk equation physical field of the local grid cell;
step 3.3), determining non-sticky and sticky physical field variables of grid body units in an overlapping area decomposed by an interface exchange area according to a parallel computing communication data structure which is pre-established to be overlapped in multiple layers;
here, the non-sticky and sticky physical field variables of the overlapping area mesh body unit decomposed using the interface exchange area established in step S101;
and 3.4) updating physical field information of the region decomposition overlapping region based on non-sticky and sticky physical field variables of the overlapping region grid body unit of the interface exchange region decomposition.
In some embodiments, each processor calculates boundary layer physical quantity characteristic quantities in the local processor according to the non-sticky physical field and the sticky physical field, records local boundary layer characteristic quantity data into a boundary layer characteristic physical quantity vector container of the wall surface grid units, and calculates global wall surface grid unit boundary layer characteristic physical quantities through MPI function global protocol for parallel calculation; the step S108 mainly includes:
The following steps are repeatedly executed for each global wall grid cell in each processor until each global wall grid cell in each processor is traversed:
step 4.1), collecting non-sticky and sticky physical quantities in the flowing physical quantities according to a grid body unit linked list of the wall grid units;
here, the global wall grid cell wi is circularly processed, the flowing physical quantity of the normal grid body cell cj of the wall grid cell wi in the wall_norm_cell [ wi ] linked list of the wall grid cell wi is processed, and non-sticky and sticky physical quantity information is collected; the normal grid body unit cj is one of data in the normal grid body unit wall_norm_cell [ wi ], and the normal grid body unit wall_norm_cell [ wi ] comprises a string of grid body units.
Step 4.2), calculating the characteristic quantity of the local boundary layer according to the non-sticky and sticky physical quantity of the normal grid body unit of the wall surface grid unit;
specifically, according to the non-sticky and sticky physical quantity of the normal grid body unit cj of the wall grid unit wi, calculating the characteristic quantity of the local boundary layer; firstly, obtaining a local velocity decomposition coordinate system to decompose viscous velocity components of a local grid, wherein the decomposition coordinate system is a normal grid body unit cj velocity unit vector a1_unit of a wall grid unit wi, and a wall unit normal vector b1_unit is calculated to obtain a transverse unit vector Calculating by using wall unit normal vector b1_unit and transverse unit vector c1_unit to obtain flow direction unit vectorThe wall unit normal vector b1_unit, the transverse unit vector c1_unit and the flow direction unit vector d1_unit form a local decomposition coordinate system, as shown in fig. 9; the calculation method of the feature quantity 1 (first feature quantity) g is as follows, a viscous flow direction velocity component u_stream=u.d1_unit=ux_unit+uy_unit+uz_unit of the viscous velocity vector u is decomposed, and a ratio of the viscous flow direction velocity component u_stream to the non-viscous velocity u_index is calculated as g=u_stream/u_index as the feature quantity 1; the characteristic quantity 1 does not need information of other grids in the boundary layer, has the characteristic of complete locality, and does not need to be associated with other grid body units during calculation; the feature quantity 2 (second feature quantity) h is based on the relevant physical quantity of the viscous physical field, specificallyWhereinNorm represents the normal direction. d (u_stream)/dnorm represents the partial derivative of the flow direction velocity in the normal direction; rho is a viscous physical field density value, and the characteristic value 2 needs to use an adjacent normal grid body unit, but because the invention constructs three-layer overlapped parallel region decomposition, the condition that the adjacent grid body unit is absent when a local processor calculates the characteristic value does not occur, and the local calculation has no difficulty; as shown in fig. 6, when calculating the feature quantity 2, a series of normal cell values such as cell a, cell B, cell C, and wall surface are used.
Step 4.3), determining characteristic points of each wall grid unit according to the local boundary layer characteristic quantity;
specifically, for the local boundary layer feature quantity of the normal grid body unit cj of the wall grid unit wi, the first feature quantity g=u_stream/u_index, when g>=g ', g' is given by the normal distance_normal_cell [ cj ]]Cj at minimum is the first feature point wall [ wi ]]-cj_character_local_1; when (when)Second characteristic quantity，/>Is given by a given value and normal distance_normal_cell [ cj ]]At least is the second feature point wall [ wi ]]-cj_character_local_2；
Step 4.4), using the parallel computing function to globally reduce the first feature point of each wall grid unit, the first normal distance between the first feature quantity (feature quantity 1) of each wall grid unit and the first normal distance, and reduce the second feature point of each wall grid unit, the second feature quantity (feature quantity 2) of each wall grid unit and the second normal distance.
First feature point wall [ wi ] of all wall grid cells by MPI function global protocol]-cj_character_local_1, and first feature quantity g and first normal distance_normal_cell [ cj_character_local_1 ]]Second feature point wall [ wi ] of all wall grid units is regulated ]-cj_character_local_2 and second normal distance h and second normal_cell [ cj_character_local_2 ]]The method comprises the steps of carrying out a first treatment on the surface of the For the first feature quantity, the global wall grid unit feature points are all g>=g ', g' is given and the first normal distance distance_normal_cell [ cj_character_local_1]The smallest grid cell, denoted wall wi]-cj_Character_1; for the second feature quantity, the global wall grid unit feature points are all，/>Is a given value and a second normal distance of normal cell cj_character_local_2]The smallest grid cell, denoted wall wi]-cj_Character_2; the main function of the MPI function global protocol operation in the step is the full collection function MPI_Allgather, see FIG. 10.
In some embodiments, the method of the preceding embodiments further comprises:
step S109, calculating other boundary layer physical quantities of the wall grid units according to the global wall grid unit boundary layer characteristic physical quantity calculated by each processor and the physical field information corresponding to the local processor; and calculating a target physical field solving value according to other boundary layer physical quantities.
It can be understood that other boundary layer physical quantity information is calculated according to the global characteristic quantity information and the local physical field calculated by each processor, and the next physical field solving value is calculated according to the newly calculated wall grid unit boundary layer physical quantity information.
This example describes an in situ distributed parallel embodiment of the method of the present invention as applied to an actual model of a fuselage assembly. The wing body combination grid is about 651 ten thousand polyhedral grid units, and single-core, double-core and 8-core distributed parallel calculation is carried out by using the boundary layer characteristic amount calculation method, wherein the calculation time is as shown in the following table 1:
table 1 in-situ distributed computing time consuming of the present embodiment
| Test appearance | Single core boundary layer feature computation time | 2 Nuclear boundary layer feature computation time | 8 kernel boundary layer feature computation time |
| Wing body assembly (651 ten thousand grid) | 7.1087 | 4.20987s | 1.86731 |
As can be seen from table 1, the boundary layer feature amount calculation method of the present embodiment has high parallelism and high calculation efficiency.
This example describes a specific embodiment of the application of the method of the invention when directed to a flat panel model. The method is applied to laminar plate boundary layer flow, the Mach number of the incoming flow is 0.3, the attack angle and the sideslip angle are both 0 degrees, the wall surface is under the adiabatic temperature condition, and the model is thatThe working condition has a theoretical relation of momentum integral, wherein +.>The Reynolds number is used to identify the flow regime, and is shown in FIG. 11 as boundary layer thickness information calculated using the method of the present invention, where the abscissa x represents the distance from the plate to the leading edge head and the ordinate θ represents the thickness of the boundary layer, and the calculated rule of the present invention is consistent with the theoretical rule.
As shown in fig. 12, an embodiment of the present invention provides a polyhedral grid in-situ distributed parallel boundary layer physical quantity acquiring apparatus, including:
the search module broadcasts the wall grid unit of each processor to the rest processors so that each processor establishes a k-d tree data structure of the wall grid unit based on the linear distance and performs adjacent point search;
the interpolation module is used for inserting the grid body unit number corresponding to each grid unit in each processor and the grid body unit of the overlapping area into a normal grid body unit linked list according to the normal distance based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points;
the acquisition module is used for sequentially solving a non-sticky Euler equation and a sticky Navistos equation according to a physical field solver by each processor to acquire physical field information of a flow field object understanding and decomposing an overlapping area based on the flow field physical solution updating area;
and the calculation module is used for calculating boundary layer physical quantity characteristic quantities in the local processors according to the physical field information of the region decomposition overlapping region and calculating global wall grid unit boundary layer characteristic physical quantities through a parallel calculation function global protocol.
In some embodiments, the apparatus further comprises a creation module for initializing a parallel computing environment, performing distributed parallel region decomposition on the input grid, creating a multi-layered overlapping parallel computing communication data structure.
In some embodiments, the search module is further specifically configured to broadcast the collected wall grid cell information to other processors by each processor; wherein the wall mesh unit information includes: the normal vector of the wall grid unit, the wall grid center coordinate, the grid point coordinates forming the wall grid and the global grid plane number; according to the position of the processor where the wall surface is located after the region decomposition, local numbering is given to the grid unit of the global wall surface; based on the linear distance, taking the face center coordinates of the global wall grid units as data, and establishing a k-d tree data structure for representing the wall grid units; and determining the nearest point of the wall grid unit to the input point based on the nearest search of the input point of the k-d tree data structure.
In some embodiments, the interpolation module is further specifically configured to repeatedly perform the following steps for each grid cell in each processor until each grid cell in each processor is traversed: searching a nearest-distance wall grid unit corresponding to each grid body unit, and recording a grid surface number of the nearest-distance wall grid unit and a linear distance between the grid body unit and the nearest-distance wall grid unit; according to the linear distance, solving the normal distance from the grid body unit to the wall grid unit, and inserting the grid body unit number into a normal grid body unit linked list of the wall grid unit according to the normal distance; starting a communication interface, decomposing distance information of grid body units in an overlapping area by an exchange area, and arranging and inserting the grid body units in the overlapping area into a normal grid body unit linked list of a target wall grid unit corresponding to the grid body units in the overlapping area according to the normal distance.
In some embodiments, the obtaining module is further specifically configured to calculate a non-stick euler equation physical field of the local grid cell by each processor; each processor calculates a viscous Navisk equation physical field for the local grid cell; according to a parallel computing communication data structure which is pre-established with multiple layers of overlapping, determining non-sticky and sticky physical field variables of grid body units in an overlapping area decomposed by an interface exchange area; and updating physical field information of the region decomposition overlapping region based on non-sticky and sticky physical field variables of the overlapping region grid body unit of the interface exchange region decomposition.
In some embodiments, the computing module is further specifically configured to repeatedly perform the following steps for each global wall grid cell in each of the processors until each global wall grid cell in each of the processors is traversed: collecting non-sticky and sticky physical quantities in the flowing physical quantities according to a grid body cell linked list of the wall grid cells; calculating the characteristic quantity of a local boundary layer according to the non-sticky physical quantity of a normal grid body unit of the wall surface grid unit; determining characteristic points of each wall grid unit according to the local boundary layer characteristic quantity; and utilizing a parallel computing function to globally reduce the first characteristic point of each wall grid unit, the first characteristic quantity and the normal distance of each wall grid unit, and reduce the second characteristic point of each wall grid unit, the second characteristic quantity and the normal distance of each wall grid unit.
In some embodiments, the calculating module is further specifically configured to calculate other boundary layer physical quantities of the wall grid units according to the global boundary layer characteristic physical quantities of the wall grid units calculated by each processor and physical field information corresponding to the local processor; and calculating a target physical field solving value according to the other boundary layer physical quantities.
In the embodiment of the present invention, the electronic device may be, but is not limited to, a personal computer (Personal Computer, PC), a notebook computer, a monitoring device, a server, and other computer devices with analysis and processing capabilities.
As an exemplary embodiment, referring to fig. 13, an electronic device 110 includes a communication interface 111, a processor 112, a memory 113, and a bus 114, the processor 112, the communication interface 111, and the memory 113 being connected by the bus 114; the memory 113 is used for storing a computer program supporting the processor 112 to execute the method, and the processor 112 is configured to execute the program stored in the memory 113.
The machine-readable storage medium referred to herein may be any electronic, magnetic, optical, or other physical storage device that can contain or store information, such as executable instructions, data, or the like. For example, a machine-readable storage medium may be: RAM (Radom Access Memory, random access memory), volatile memory, non-volatile memory, flash memory, a storage drive (e.g., hard drive), any type of storage disk (e.g., optical disk, dvd, etc.), or a similar storage medium, or a combination thereof.
The non-volatile medium may be a non-volatile memory, a flash memory, a storage drive (e.g., hard drive), any type of storage disk (e.g., optical disk, dvd, etc.), or a similar non-volatile storage medium, or a combination thereof.
It can be understood that the specific operation method of each functional module in this embodiment may refer to the detailed description of the corresponding steps in the above method embodiment, and the detailed description is not repeated here.
The computer readable storage medium provided by the embodiments of the present invention stores a computer program, where the computer program code may implement the method described in any of the foregoing embodiments when executed, and the specific implementation may refer to the method embodiment and will not be described herein.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described system and apparatus may refer to corresponding procedures in the foregoing method embodiments, which are not described herein again.
In addition, in the description of embodiments of the present invention, unless explicitly stated and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
In the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above examples are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention, but it should be understood by those skilled in the art that the present invention is not limited thereto, and that the present invention is described in detail with reference to the foregoing examples: any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or perform equivalent substitution of some of the technical features, while remaining within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (9)
1. A polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method is characterized by comprising the following steps of:
broadcasting the wall grid unit of each processor to the rest processors, so that each processor establishes a k-d tree data structure of the wall grid unit based on the linear distance, and performs adjacent point search;
based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points, inserting the grid body unit number corresponding to each grid unit in each processor and the grid body unit of the overlapping area into a normal grid body unit linked list according to the normal distance;
each processor sequentially solves a non-sticky Euler equation and a sticky Navistos equation according to a physical field solver to obtain flow field object understanding and update physical field information of a region decomposition overlapping region based on the physical solution of the flow field;
each processor calculates boundary layer physical quantity characteristic quantity in the local processor according to the physical field information of the region decomposition overlapping region, and calculates global wall grid unit boundary layer characteristic physical quantity through a parallel calculation function global protocol;
before broadcasting the wall grid cells of each processor to the remaining processors to cause the each processor to build a k-d tree data structure of the wall grid cells based on the straight line distance and to perform the step of searching for neighboring points, the method further comprises:
Initializing a parallel computing environment, carrying out distributed parallel region decomposition on grids of an input wing body assembly model, and creating a multi-layer overlapped parallel computing communication data structure.
2. The method of claim 1, wherein broadcasting the wall grid cells of each processor to the remaining processors such that each processor builds a k-d tree data structure of the wall grid cells based on the linear distance and performs the neighbor search comprises:
each processor broadcasts the collected wall grid cell information to other processors; wherein the wall mesh unit information includes: the normal vector of the wall grid unit, the wall grid center coordinate, the grid point coordinates forming the wall grid and the global grid plane number;
according to the position of the processor where the wall surface is located after the region decomposition, local numbering is given to the grid unit of the global wall surface;
based on the linear distance, taking the face center coordinates of the global wall grid units as data, and establishing a k-d tree data structure for representing the wall grid units;
and determining the nearest point of the wall grid unit to the input point based on the nearest search of the input point of the k-d tree data structure.
3. The method of claim 1, wherein the step of inserting the grid body cell number and the overlap area grid body cell corresponding to each grid cell in each processor into the normal grid body cell linked list according to the normal distance based on the nearest-distance wall grid cell of the grid body cell determined by the adjacent point comprises:
the following steps are repeatedly performed for each grid cell in each processor until each of the grid cells in each processor is traversed:
searching a nearest-distance wall grid unit corresponding to each grid body unit, and recording a grid surface number of the nearest-distance wall grid unit and a linear distance between the grid body unit and the nearest-distance wall grid unit;
according to the linear distance, solving the normal distance from the grid body unit to the wall grid unit, and inserting the grid body unit number into a normal grid body unit linked list of the wall grid unit according to the normal distance;
starting a communication interface, decomposing distance information of grid body units in an overlapping area by an exchange area, and arranging and inserting the grid body units in the overlapping area into a normal grid body unit linked list of a target wall grid unit corresponding to the grid body units in the overlapping area according to the normal distance.
4. The method of claim 1, wherein each processor sequentially solves a non-viscous euler equation and a viscous navistokes equation according to a physical field solver, and wherein the step of obtaining a flow field object understanding and updating physical field information of a region-resolved overlap region based on the flow field physical solution comprises:
each processor calculates a physical field of a non-stick Euler equation of the local grid unit;
each processor calculates a viscous Navisk equation physical field for the local grid cell;
according to a parallel computing communication data structure which is pre-established with multiple layers of overlapping, determining non-sticky and sticky physical field variables of grid body units in an overlapping area decomposed by an interface exchange area;
and updating physical field information of the region decomposition overlapping region based on non-sticky and sticky physical field variables of the overlapping region grid body unit of the interface exchange region decomposition.
5. The method according to claim 1, wherein the step of each processor calculating boundary layer physical quantity characteristic quantities in the local processor according to the physical field information of the region decomposition overlap region, and calculating global wall grid cell boundary layer characteristic physical quantities by a parallel calculation function global protocol comprises:
Repeating the following steps for each global wall grid cell in each processor until each global wall grid cell in each processor is traversed:
collecting non-sticky and sticky physical quantities in the flowing physical quantities according to a grid body cell linked list of the wall grid cells;
calculating the characteristic quantity of a local boundary layer according to the non-sticky physical quantity of a normal grid body unit of the wall surface grid unit;
determining characteristic points of each wall grid unit according to the local boundary layer characteristic quantity;
and utilizing a parallel computing function to globally reduce the first characteristic point of each wall grid unit, the first characteristic quantity and the normal distance of each wall grid unit, and reduce the second characteristic point of each wall grid unit, the second characteristic quantity and the normal distance of each wall grid unit.
6. The method according to claim 2, wherein the method further comprises:
the global boundary layer characteristic physical quantity of the wall grid unit calculated by each processor and the physical field information corresponding to the local processor calculate other boundary layer physical quantities of the wall grid unit;
And calculating a target physical field solving value according to the other boundary layer physical quantities.
7. The polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition device is characterized by comprising:
the search module broadcasts the wall grid unit of each processor to the rest processors so that each processor establishes a k-d tree data structure of the wall grid unit based on the linear distance and performs adjacent point search;
the interpolation module is used for inserting the grid body unit number corresponding to each grid unit in each processor and the grid body unit of the overlapping area into a normal grid body unit linked list according to the normal distance based on the nearest distance wall grid unit of the grid body unit determined by the adjacent points;
the acquisition module is used for sequentially solving a non-sticky Euler equation and a sticky Navistos equation according to a physical field solver by each processor to acquire physical field information of a flow field object understanding and decomposing an overlapping area based on the flow field physical solution updating area;
the computing module is used for computing boundary layer physical quantity characteristic quantities in the local processor according to the physical field information of the region decomposition overlapping region, and computing global wall grid unit boundary layer characteristic physical quantities through a parallel computing function global protocol;
Before broadcasting the wall grid unit of each processor to the rest of the processors, so that each processor builds a k-d tree data structure of the wall grid unit based on the linear distance, and performs the adjacent point search, the method further comprises:
initializing a parallel computing environment, carrying out distributed parallel region decomposition on grids of an input wing body assembly model, and creating a multi-layer overlapped parallel computing communication data structure.
8. An electronic device comprising a memory, a processor and a program stored on the memory and capable of running on the processor, the processor implementing the method of any one of claims 1 to 6 when executing the program.
9. A computer readable storage medium, characterized in that the computer program is stored in the readable storage medium, which computer program, when executed, implements the method of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310849613.1A CN116579269B (en) | 2023-07-12 | 2023-07-12 | Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310849613.1A CN116579269B (en) | 2023-07-12 | 2023-07-12 | Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116579269A CN116579269A (en) | 2023-08-11 |
| CN116579269B true CN116579269B (en) | 2023-09-26 |
Family
ID=87541672
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310849613.1A Active CN116579269B (en) | 2023-07-12 | 2023-07-12 | Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116579269B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109948002A (en) * | 2019-03-14 | 2019-06-28 | 空气动力学国家重点实验室 | The nearest wall surface of unstrctured grid based on balance KD tree is apart from method for solving |
| CN110633149A (en) * | 2019-09-10 | 2019-12-31 | 中国人民解放军国防科技大学 | Parallel load balancing method for balancing calculation amount of unstructured grid unit |
| CN115879231A (en) * | 2023-02-24 | 2023-03-31 | 北京卓鸷科技有限责任公司 | Method and device for acquiring flow transition information of micro aircraft |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8259104B2 (en) * | 2008-08-28 | 2012-09-04 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Domain decomposition by the advancing-partition method for parallel unstructured grid generation |
-
2023
- 2023-07-12 CN CN202310849613.1A patent/CN116579269B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109948002A (en) * | 2019-03-14 | 2019-06-28 | 空气动力学国家重点实验室 | The nearest wall surface of unstrctured grid based on balance KD tree is apart from method for solving |
| CN110633149A (en) * | 2019-09-10 | 2019-12-31 | 中国人民解放军国防科技大学 | Parallel load balancing method for balancing calculation amount of unstructured grid unit |
| CN115879231A (en) * | 2023-02-24 | 2023-03-31 | 北京卓鸷科技有限责任公司 | Method and device for acquiring flow transition information of micro aircraft |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116579269A (en) | 2023-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bhatnagar et al. | Prediction of aerodynamic flow fields using convolutional neural networks | |
| Secco et al. | Efficient mesh generation and deformation for aerodynamic shape optimization | |
| Nielsen et al. | An implicit, exact dual adjoint solution method for turbulent flows on unstructured grids | |
| Luo et al. | A reconstructed discontinuous Galerkin method based on a hierarchical WENO reconstruction for compressible flows on tetrahedral grids | |
| Weinkauf et al. | Stable feature flow fields | |
| CN112818573B (en) | Method for acquiring boundary layer non-local variable information for unstructured grid | |
| Kumar et al. | An efficient immersed boundary algorithm for simulation of flows in curved and moving geometries | |
| Kröger et al. | CAD-free hydrodynamic optimisation using consistent kernel-based sensitivity filtering | |
| Leifsson et al. | Optimal shape design of multi-element trawl-doors using local surrogate models | |
| Benamara et al. | Adaptive infill sampling criterion for multi-fidelity optimization based on Gappy-POD: Application to the flight domain study of a transonic airfoil | |
| Peng et al. | An automatic isotropic/anisotropic hybrid grid generation technique for viscous flow simulations based on an artificial neural network | |
| Iyengar et al. | Nonlinear reduced order modeling using domain decomposition | |
| Durasov et al. | Debosh: Deep bayesian shape optimization | |
| Zhang et al. | Efficient aerodynamic shape optimization of the hypersonic lifting body based on free form deformation technique | |
| Xu et al. | G2 surface modeling using minimal mean-curvature-variation flow | |
| Kaya et al. | Discrete adjoint-based aerodynamic shape optimization framework for natural laminar flows | |
| CN116579269B (en) | Polyhedral grid in-situ distributed parallel boundary layer physical quantity acquisition method and device | |
| Toal et al. | Geometric filtration using POD for aerodynamic design optimization | |
| Strofylas et al. | An agglomeration strategy for accelerating RBF-based mesh deformation | |
| Turner et al. | Automatic generation of 3D unstructured high-order curvilinear meshes | |
| Coirier et al. | A Cartesian, cell-based approach for adaptively-refined solutions of the Euler and Navier-Stokes equations | |
| Frink | Three-dimensional upwind scheme for solving the Euler equations on unstructured tetrahedral grids | |
| Lokatt et al. | Finite-volume scheme for the solution of integral boundary layer equations | |
| Porter et al. | Three Dimensional Surface Redefinition Method for Computational Ice Accretion Solvers | |
| Mufti et al. | Shock wave prediction in transonic flow fields using domain-informed probabilistic deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |